2008 giraud kellen senior thesis
Post on 14-Apr-2018
221 Views
Preview:
TRANSCRIPT
-
7/29/2019 2008 Giraud Kellen Senior Thesis
1/79
SIMULATION AND MANUFACTURE OF A QUADRUPOLE MASS FILTER
FOR A 7BE ION PLASMA
by
Kellen M. Giraud
A senior thesis submitted to the faculty of
Brigham Young University
in partial fulfillment of the requirements for the degree of
Bachelor of Science
Department of Physics and Astronomy
Brigham Young University
April 2008
-
7/29/2019 2008 Giraud Kellen Senior Thesis
2/79
Copyright c 2008 Kellen M. Giraud
All Rights Reserved
-
7/29/2019 2008 Giraud Kellen Senior Thesis
3/79
BRIGHAM YOUNG UNIVERSITY
DEPARTMENT APPROVAL
of a senior thesis submitted by
Kellen M. Giraud
This thesis has been reviewed by the research advisor, research coordinator,and department chair and has been found to be satisfactory.
Date Grant Hart, Advisor
Date Eric Hintz, Research Coordinator
Date Ross Spencer, Chair
-
7/29/2019 2008 Giraud Kellen Senior Thesis
4/79
ABSTRACT
SIMULATION AND MANUFACTURE OF A QUADRUPOLE MASS FILTER
FOR A 7BE ION PLASMA
Kellen M. Giraud
Department of Physics and Astronomy
Bachelor of Science
A quadrupole mass filter (QMF) has been simulated and manufactured. The
quadrupole is part of an ion trap system that will isolate a plasma of ionized
beryllium-7 (7Be) for months at a time. Through this experiment, the half-
life of ionized 7Be will be measured. The function of the QMF is to filter
out unwanted ions and allow wanted ions to penetrate into the region of high
magnetic field. Voltages are applied to the conducting rods of the QMF based
on the stability curve generated by the Mathieu equations. The motion of7Be
ions in an ideal QMF has been simulated. Our QMF has been altered from
the ideal QMF, and the motion of ions in the modified QMF has also been
simulated. A quadrupole was built based on the results of the simulations,
and was installed into the experimental system. Tests on the manufactured
quadrupole have indicated that significant changes must yet be made for the
experimental QMF to match the simulations.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
5/79
ACKNOWLEDGMENTS
Many people made this research possible and contributed to my project in
ways I cannot possibly enumerate.
I am grateful for my wife, Marie, for supporting me in the many hours
spent working on my project. I thank all of my family, who have been very
supportive and encouraging of my work. I especially thank my brother Gavin
for leading the way in studying physics.
I am grateful for my advisors Dr. Hart and Dr. Peterson for the time and
energy they have spent on my behalf. The rest of the plasma research group
has also been a great support to me. David Olson especially contributed to
this project in helping to design the support structure for the quadrupole.
I thank the BYU Office of Research and Creative Activities and the De-
partment of Physics and Astronomy for funding to work on this project.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
6/79
Contents
Table of Contents vi
List of Figures viii
1 Introduction 11.1 Why Beryllium-7 Is Interesting . . . . . . . . . . . . . . . . . . . . . 11.2 The Quadrupole Mass Filter . . . . . . . . . . . . . . . . . . . . . . . 51.3 Summary of Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.4 Summary of Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 7
2 Methods 9
2.1 The Ideal Quadrupole . . . . . . . . . . . . . . . . . . . . . . . . . . 92.2 The Modified Quadrupole . . . . . . . . . . . . . . . . . . . . . . . . 132.3 Computer Modeling of the Quadrupole . . . . . . . . . . . . . . . . . 172.4 The Magnetic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5 The Modified Stability Curve . . . . . . . . . . . . . . . . . . . . . . 252.6 Quadrupole Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . 26
3 Results and Conclusions 28
3.1 Comparison of Simulation and Experiment . . . . . . . . . . . . . . . 283.2 Materials in the Quadrupole . . . . . . . . . . . . . . . . . . . . . . . 313.3 The Effect of the Magnetic Field . . . . . . . . . . . . . . . . . . . . 333.4 Reduction of Fringing Fields . . . . . . . . . . . . . . . . . . . . . . . 343.5 Potential QMF Modifications . . . . . . . . . . . . . . . . . . . . . . 363.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Bibliography 39
A The Mathieu Equation 42
B Matlab Code 46
B.1 Solving the Voltage Grid . . . . . . . . . . . . . . . . . . . . . . . . . 46B.2 Solving the Electric Fields . . . . . . . . . . . . . . . . . . . . . . . . 51B.3 The Magnetic Field Values . . . . . . . . . . . . . . . . . . . . . . . . 52
vi
-
7/29/2019 2008 Giraud Kellen Senior Thesis
7/79
CONTENTS vii
B.4 The Magnetic Field Grid . . . . . . . . . . . . . . . . . . . . . . . . . 54B.5 The Mover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55B.6 Solving the Ion Motion . . . . . . . . . . . . . . . . . . . . . . . . . . 56
C Machining Drawings 62
D Pictures 67
Index 67
-
7/29/2019 2008 Giraud Kellen Senior Thesis
8/79
List of Figures
1.1 Ion trap assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.2 Quadrupole schematic . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 Stability curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Ideal quadrupole ion trajectories . . . . . . . . . . . . . . . . . . . . . 122.3 Comparison of stability curves for ions in the experiment . . . . . . . 142.4 Comparison of hyperbolic and best-fit cylindrical surfaces . . . . . . . 152.5 Voltage contours for three quadrupole geometries . . . . . . . . . . . 162.6 Electric fields in the modified quadrupole . . . . . . . . . . . . . . . . 182.7 Conceptual picture of magnetic field lines . . . . . . . . . . . . . . . . 202.8 Magnetic field in the quadrupole . . . . . . . . . . . . . . . . . . . . . 212.9 Resonance in the QMF . . . . . . . . . . . . . . . . . . . . . . . . . . 242.10 The modified stability curve . . . . . . . . . . . . . . . . . . . . . . . 262.11 Design for QMF and support system . . . . . . . . . . . . . . . . . . 27
3.1 Voltages produced by a charged plate . . . . . . . . . . . . . . . . . . 323.2 Charge collected by the QMF with and without magnetic field . . . . 343.3 Ion trajectories in a fringing field . . . . . . . . . . . . . . . . . . . . 35
C.1 Support plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62C.2 Ion source bracket 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 63C.3 Ion source bracket 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 64C.4 Ground plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65C.5 Far end support of the QMF . . . . . . . . . . . . . . . . . . . . . . . 66
D.1 The experimental system . . . . . . . . . . . . . . . . . . . . . . . . . 68
D.2 Side view of the QMF . . . . . . . . . . . . . . . . . . . . . . . . . . 68D.3 Close up view of mounting system . . . . . . . . . . . . . . . . . . . . 69
viii
-
7/29/2019 2008 Giraud Kellen Senior Thesis
9/79
Chapter 1
Introduction
1.1 Why Beryllium-7 Is Interesting
Beryllium, the fourth element in the periodic table, occupies a place of interest as one
of the lightest metals. It is found in many mineral compounds. Beryllium has only
one stable isotope, 9Be. Unstable isotopes of beryllium include 10Be, with a half-life
of about 1.5 million years and 7Be, with a half-life of about 53 days. The remaining
isotopes of beryllium have half-lives in the range of seconds or a small fraction of a
second. The isotope of interest in this paper is 7Be.
7Be is found in various locations. It is part of the fusion cycle in stellar cores.[1] It
is also created in cosmic ray reactions in earths atmosphere. From the region of peak
production at an altitude of about 20 km, 7Be makes its way to earths surface.[2]
7Be is of interest in studies ranging from soil erosion[3] to sheep droppings.[4]
In every study dealing with 7Be, its half-life is an important factor. However, 7Be
does not have a straightforward constant half-life like most isotopes. The decay rate
of7Be significantly depends on its chemical environment. This is because 7Be decays
1
-
7/29/2019 2008 Giraud Kellen Senior Thesis
10/79
CHAPTER 1. INTRODUCTION 2
Table 1.1: Half-life of7Be in various environments
Environment Half-life (days) Percent change
Accepted value 53.29 0
Buckyball 52.68 -1.14%
BeO 54.226 +1.76%
Be(OH)2 at 442 kbar 52.884 -0.76%
exclusively by electron capture,
7
Be + e
7
Li + (1.1)
into lithium-7 and a neutrino. Because the nucleus must capture an electron to decay,
the decay rate is dependent on electron density at the nucleus. In most elements that
have isotopes that decay exclusively by electron capture, it is difficult to influence
electron density at the nucleus because these elements have high atomic numbers. In
elements with higher atomic numbers, electrons on the outer shells of the atom shield
inner electrons from any chemical changes that might influence electron density atthe nucleus. However, since 7Be only has four electrons, the chemical environment of
7Be can significantly influence the electron density at the nucleus, and thus influence
the availability of electrons for capture to the nucleus. For this reason, the chemical
environment of 7Be can influence its half-life.
Various research groups exploring the variable half-life of 7Be in a variety of sit-
uations have observed a 1 to 2 % change from the usual half-life of about 53 days,
as shown in Table 1.1. Previously used methods to observe a change in decay rate
of 7Be include placing 7Be inside a buckyball (C60),[5] placing it in the compound
BeO,[6] and exposing the compound Be(OH)2 to extreme pressures.[7]
In spite of success in manipulating the half-life of 7Be, no group has measured
the half-life of 7Be in an ionized state. In such a state, the half-life is expected to
-
7/29/2019 2008 Giraud Kellen Senior Thesis
11/79
CHAPTER 1. INTRODUCTION 3
change by as much as 5%. This is a significant change in any half-life considering the
great accuracy with which half-life values are usually measured and reported. Among
all the possible experiments that could be done to measure the half-life of 7Be in a
variety of environments, it is particularly important to determine the half-life in an
ionized state. This is because 7Be is ionized in many cases in nature, including the
places of most interest to science, in earths upper atmosphere and in nuclear fusion
cycles in stellar cores.
Measuring the decay rate of ionized 7Be requires a very specific set of conditions.
First, the ions must be isolated in a system in which they can remain in a plasmastate. Second, they must be isolated for an amount of time comparable to the half-life
of 53 days. Third, there must be enough ions in the system for statistics to be taken
that determine the half-life to the desired accuracy. Since decay is a random process,
gathering enough statistics is important. Combining these conditions, there must be
a system that holds an isolated plasma of about 109 ions of7Be for months at a time.
The plasma research group at BYU has devised a process to measure the half-life
of ionized 7Be. The process will take place in the system shown schematically in Fig.
1.1. For the process to be initiated, first a sample must be prepared for placement in
the system. To obtain a sample containing 7Be, a small boron carbide (B4C) disk of
radius 0.75 in. is bombarded with protons in a Van de Graff accelerator. The protons
have an average energy of approximately 400 kV. A trace amount of7Be is produced
in nuclear reactions,
10
B + p+
7
Be + (1.2)
When high-energy protons strike boron-10, the most abundant of borons isotopes,
an alpha particle is ejected from the nucleus and beryllium-7 remains in the sample.
Next, the boron carbide sample infused with traces of 7Be is placed in a modified
metal vapor vacuum arc (MEVVA) ion source[8] that ejects ions into a quadrupole
-
7/29/2019 2008 Giraud Kellen Senior Thesis
12/79
CHAPTER 1. INTRODUCTION 4
Ion Source
Ground Plate
Quadrupole
Ion Trap
Magnet CoilsVacuum Pump
ChargeCollector
Figure 1.1: A simplified schematic of the ion trapping system. See appendix D for an
actual picture of the system.
mass filter (QMF). The quadrupole filters out all ions except 7Be which is then chan-
neled into a Malmberg-Penning ion trap.[9] Once the ions are in the trap, they will
be kept isolated inside for up to two months. During this time, measurements of
the ratio of7
Be to7
Li will be performed using the Fourier Transform Ion CyclotronResonance Mass Spectrometry (FTICR-MS) method.
The process of firing the ion source and channeling ions into the trap can be
repeated as many times as necessary to build up a sufficient number of ions in the
trap. In order to calibrate the experiment, charge collectors are located at the far
end of the system. These charge collectors measure the amount of charge on metal
disks of various radii so that we can tell the radial profile of ions in our system. That
is, we will know the amount of ions in the center of the trap and near the edges.
After the boron carbide sample is placed in the MEVVA ion source, the entire
system is pumped down to ultra-high vacuum levels. The quadrupole bridges the gap
from where the ion plasma is produced in the ion source to where it is isolated in the
-
7/29/2019 2008 Giraud Kellen Senior Thesis
13/79
CHAPTER 1. INTRODUCTION 5
ion trap.
1.2 The Quadrupole Mass Filter
In our experiment to determine the half-life of ionized 7Be, a quadrupole mass filter
will be used to isolate and channel 7Be ions into the ion trap. Because a boron carbide
sample is used in the experiment, the ions produced by the ion source will have an
overwhelming concentration of boron-10, boron-11, and carbon-12. We predict that
7Be ions will actually make up only about one in 105 ions from the ion source. The
quadrupole must be able to filter out everything except the small number of 7Be ions.
A quadrupole mass filter is a device that selects ions with a specific charge-to-mass
ratio or range of charge-to-mass ratios. A QMF consists of four conducting rods to
which a specific combination of direct and alternating voltages are applied. As shown
in Fig. 1.2, opposing voltages are applied to adjacent rods. At any particular time,
if a certain rod is at a voltage V, the two rods nearest it are at a voltage negative
V. These voltages generate electric fields around the rods, which produce forces on
charged particles in the region. The voltages on the rods determine the forces that
drive the motion of ions in the QMF. Thus, a certain combination of voltages drives
out unwanted ions while allowing wanted ions to pass through the quadrupole.
The quadrupole also serves the purpose of guiding the 7Be ions into a region of
strong magnetic field. The Malmberg-Penning trap confines the ions by means of a
magnetic field that reaches a peak strength of 0.43 T. However, the ion source liesoutside the region of the magnetic coils. The field strength at the ion source is only
about 4103 T, about one hundredth of the peak value. Charged particles that enter
a region of high magnetic field from a region of low magnetic field often encounter the
magnetic mirror effect.[10] The quadrupole serves the function of breaking this effect
-
7/29/2019 2008 Giraud Kellen Senior Thesis
14/79
CHAPTER 1. INTRODUCTION 6
U+Vcos(t) -[U+Vcos(t)]
-[U+Vcos(t)]U+Vcos(t)
Figure 1.2: Voltages applied to the rods of a quadrupole that allow for selection of
ions with a specified charge-to-mass ratio. U is direct voltage while V is alternating
voltage.
and allowing ions to penetrate the high magnetic field. In our experiment, however,
we have found that this effect is of minimal importance, as will be shown.
1.3 Summary of Methods
I have modeled the motion of ions in a quadrupole mass filter. This included solving
the Mathieu equations to simulate motion in an ideal quadrupole, solving for electric
potential in grids using Poissons equation in two and three dimensions, and solving
for ion motion in electric and magnetic fields using a differential equation solver. This
involved many hours of coding in Matlab (see appendix B for Matlab code). In
chapter two I outline my work on computer simulations of the QMF.
Using the results of the simulations, I helped to design the QMF and support
structure for the experimental system. Graduate student David Olson and I designed
the parts for the quadrupole and support structure in SolidWorks, and the pieces were
-
7/29/2019 2008 Giraud Kellen Senior Thesis
15/79
CHAPTER 1. INTRODUCTION 7
manufactured in October 2007 (see appendix C for machining drawings of parts).
Since the quadrupole was put together and placed in the experimental system, I
have been involved in many tests run for the purpose of characterizing the behavior
of the QMF. The ion source has been fired hundreds of times with the voltages of the
quadrupole and strength of the magnetic field set to various values while data has
been taken for analysis.
1.4 Summary of Conclusions
While much data has been taken to analyze the motion of ions in the fields of our
quadrupole, there remains much work to be done before 7Be can be successfully
isolated in the ion trap. In some aspects, the physical QMF produces results that
parallel the computer simulations, while in other respects, the behavior of the ions in
the QMF is not yet well understood. There are many complexities to the experiment
that complicate the analysis of the quadrupole. In chapter three I discuss the data
that have been generated by experimentation, the successes and shortcomings of my
simulations to predict the behavior of the experimental quadrupole, and possible
improvements that can be made to the QMF.
One obvious result from tests on the quadrupole was that the magnetic field of the
ion trap has a significantly greater influence on the motion of ions in the quadrupole
than was expected based on computer simulations. We found that when the magnetic
field is on, an average of 100 times more charge accumulates at the charge collectorsthan when the field is off. This probably means that the magnetic field focuses ions
in the system significantly more than was expected.
We also found a problem in our original design of the QMF. In our tests, we
found that after only a few tests of the quadrupole the data became completely
-
7/29/2019 2008 Giraud Kellen Senior Thesis
16/79
CHAPTER 1. INTRODUCTION 8
unpredictable and unrepeatable. This was attributed to the fact that a layer of the
insulator aluminum oxide built up on all aluminum surfaces. Thus, ions from the
ion source striking the aluminum oxide surface remained on the insulator as a surface
charge. This distorted the fields of the QMF and prevented ions from getting through
the quadrupole. In order to fix this problem, we replaced the some of the aluminum
parts of the quadrupole with copper parts. Copper oxide is a semiconductor which
allows any charge buildup on the quadrupole surfaces to discharge in a reasonable
amount of time.
As of current testing, the quadrupole shows little evidence of being able to filterout ions of unwanted mass to charge ratio to the degree of precision required for
our experiment. There are many possible solutions to this problem. Many other
groups [15-19] have worked with QMFs to enhance the filtration properties by mod-
ifications of the QMF design. It may be possible to employ modifications to enable
our quadrupole to function as required in the experiment to measure the half-life of
ionized 7Be.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
17/79
Chapter 2
Methods
2.1 The Ideal Quadrupole
This chapter outlines my methods for simulating and manufacturing the quadrupole
mass filter (QMF). The quadrupole is an important piece in the system, shown in Fig.
1.1, that will confine a 7Be plasma for months at a time. Through this experiment,
the half-life of ionized 7Be will be measured.
Building a quadrupole for our experiment that selectively allows 7Be into the
ion trap required starting from basic quadrupole theory. This section analyzes the
ideal QMF. The ideal QMF is very different than the quadrupole used in our experi-
ment; however, it outlines the basic theory required for understanding any QMF. All
quadrupoles use conducting surfaces with voltages applied in a specific way to achieve
mass filtration, as show in in Fig. 1.2. The most essential ingredient in understanding
any QMF is the stability curve plot, shown in Fig. 2.1. A basic understanding of this
plot comes from the ideal quadrupole.
Conceptually, any quadrupole functions as a mass filter because ions with greater
mass than the target ion are ejected in the y direction while ions with lesser mass
9
-
7/29/2019 2008 Giraud Kellen Senior Thesis
18/79
CHAPTER 2. METHODS 10
0 0.2 0.4 0.6 0.8 1 1.20
0.05
0.1
0.15
0.2
0.25
q
a
Unstable
in xdirectionUnstable
in ydirection
Stable
Figure 2.1: Stability curves from the Mathieu equations. The left curve separates
stable and unstable motion in the y direction, while the right curve separates stable
and unstable motion in the x direction. The stars are positions in stability space of
the ions plotted in Fig. 2.2.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
19/79
CHAPTER 2. METHODS 11
are ejected in the x direction. Larger ions are ejected because they have a greater
inertia and do not respond to the oscillating fields of the quadrupole. Smaller ions are
ejected because they respond to the oscillating fields so much that they are thrown into
unstable motion. Stable ions have a balance between the static force caused by the
DC voltage and the ponderomotive force caused by the AC voltage. A ponderomotive
force occurs in rapidly oscillating fields and pushes particles steadily. In effect, stable
ions have a balance between electrostatic forces and ponderomotive forces that enables
them to stay in the QMF despite rapid oscillations.
The ideal quadrupole is the simplest way to examine the behavior of the QMFtheoretically. Physically, the ideal quadrupole is an arrangement of four conducting
hyperbolic surfaces that extend out to infinity. This geometry, of course without
extending out to infinity, is shown in Fig. 2.2. In such a geometry, the motion
of charged particles in the quadrupole can be described by a differential equation
known as the Mathieu equation.[11] The details of the Mathieu equation analysis are
provided in Appendix A. The important result of this analysis is the stability curve
diagram.
The stability curves, shown in Fig. 2.1, describe the conditions under which a
particle is stable in a QMF. They are typically plotted based on parameters q and
a. Here,
q =4eV
mr202
(2.1)
and
a = 8eUmr20
2(2.2)
In these equations, e is the charge of the ion, m is the mass of the ion, and r0
is the inner radius of the quadrupole (the distance from the center of the QMF to
the closest surface of one of the poles). U is the value of the applied direct current
(DC) voltage and V is the value of the applied zero-to-peak alternating current (AC)
-
7/29/2019 2008 Giraud Kellen Senior Thesis
20/79
CHAPTER 2. METHODS 12
voltage. The value of is the frequency at which the alternating current oscillates.
The value of r0 is fixed by the geometry of the quadrupole and e and m are fixed
according to the experiment parameters. The values of U, V, and can be adjusted
and applied to the quadrupole.
(a) Stable motion. Here, q = 0.7
and a = 0.15.
(b) Unstable motion in y direction.
Here, q = 0.5 and a = 0.15.
(c) Unstable motion in x direction.
Here, q = 0.78 and a = 0.15.
Figure 2.2: Ion trajectories for ions in the ideal quadrupole with various values for
the parameters q and a. These values determine where the ion lies in stability space,
and thus whether the ion will have motion that stays fixed in some boundary or that
grows exponentially away from the origin.
Fundamentally, the stability curves tell us what DC and AC voltages to apply to
the quadrupole to obtain stable ion motion. If we define the values of e, m, r0, and
-
7/29/2019 2008 Giraud Kellen Senior Thesis
21/79
CHAPTER 2. METHODS 13
in an experiment, then the parameters q and a give a direct relationship between
DC and AC voltage.
The stability curves separate regions of stability and instability in the quadrupole.
The stable region is the area under both curves in Fig. 2.1. Ions with parameters
that produce values of q and a that fit in this stability region can remain indefinitely
in the quadrupole, as shown in the trajectory in Fig. 2.2(a). The curve on the left
in Fig. 2.1 is the line separating stable motion from motion that is unstable in the y
direction, as shown in the trajectory in Fig. 2.2(b). The curve on the right is the line
separating stable motion from motion that is unstable in the x direction, as shown inthe trajectory in Fig. 2.2(c).
In our experiment, e and m are the values for singly ionized 7Be. That is, e = +1
electron charge or 1.602 1019 Coulombs and m = 7.017 amu or 1.165 1026 kg.
We will be using AC voltage with a frequency in the range of 500-600 kHz. When
we apply these parameters to the stability curves, the stability curve plot becomes
a more concrete way to understand how we need to operate the quadrupole. Figure
2.3 shows the stability curve for 7Be as compared to the stability curves for boron-10,
boron-11, and carbon-12 as a function of AC and DC voltage.
2.2 The Modified Quadrupole
Ideally, hyperbolic surfaces are used in a quadrupole. Hyperbolic surfaces yield the
greatest mass resolution for the least amount of power.[12] However, cylindrical sur-faces typically suffice for most applications and are widely used. Metal rods are much
easier to obtain than hyperbolic surfaces.
It is often convenient to use cylindrical surfaces that best approximate hyperbolic
surfaces. Rods with a radius equal to 1.1487 times the inner radius of the QMF
-
7/29/2019 2008 Giraud Kellen Senior Thesis
22/79
CHAPTER 2. METHODS 14
Figure 2.3: Stability curves for 7Be, 10B, 11B, and 12C as a function of AC and DC
voltage. The values here are for a quadrupole with r0 = 2 cm and a frequency of 550
kHz. The white regions are regions stable for only one mass, while the darkest region
is stable for all masses, with regions stable for two and three masses in between. We
can operate the quadrupole at the voltages indicated by the star so that only 7Be is
stable.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
23/79
CHAPTER 2. METHODS 15
r0
r
Figure 2.4: The hyperbolic curves that define the ideal quadrupole are overlaid with
the best-fit cylindrical curves. The value r0 is the distance from the origin to the
point of closest approach for the hyperbolic or cylindrical surfaces. The value r is the
radius of the cylindrical rod.
best approximate hyperbolic surfaces.[13] As shown in Fig. 2.4, this radius yields
cylindrical surfaces that very closely match the hyperbolic surfaces in the region
closest to the origin.
In our experiment, we modified the quadrupole further by reducing the ratio of
rod size to inner radius below the best-fit ratio. Instead of the best fit r = 1.1487r0,
we used r = 0.5r0. Specifically, our QMF has an inner radius of 2.0475 cm and a rod
radius of 0.9525 cm. This additional modification was made to enhance the vacuum
capability of the ion-trapping system as a whole.
Our experiment requires vacuum levels on the order of 109 torr. In this range
of vacuum, the density of gas becomes so low that gas molecules move by molecular
flow. That is, they randomly bounce off of walls until eventually entering and being
trapped by the vacuum pump. As shown in Fig. 1.1, the quadrupole is located
between the ion trap and the vacuum pump. In order for gas molecules in molecular
flow to get to the vacuum pump, they must pass through the quadrupole. Therefore,
the quadrupole rod size was reduced to allow for the maximum amount of free space
-
7/29/2019 2008 Giraud Kellen Senior Thesis
24/79
CHAPTER 2. METHODS 16
0.05 0 0.050.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
(a) Ideal quadrupole
0.05 0 0.050.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
(b) Best-fit quadrupole
0.05 0 0.050.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
(c) Modified quadrupole
Figure 2.5: Equipotential contours for the ideal quadrupole, the best-fit cylindrical
quadrupole, and the modified cylindrical quadrupole. Dark lines represent positive
voltage, while light lines represent negative voltage. The lines crossing at the origin
are at 0 V for all three geometries.
for particles to get to the vacuum pump.
Figure 2.5 is a comparison of voltage contours in three geometries of QMFs. The
voltage values near the origin are most important for ion mass selection. As can be
seen here, the voltage values near the origin for the modified quadrupole, Fig. 2.5(c),
are very similar to the voltages in the best-fit quadrupole, Fig. 2.5(b), and the ideal
quadrupole, Fig. 2.5(a).
While it is convenient in our experiment to make the change from the ideal
-
7/29/2019 2008 Giraud Kellen Senior Thesis
25/79
CHAPTER 2. METHODS 17
quadrupole to the cylindrical quadrupole with rods smaller than the best fit size,
this change modifies the function of the quadrupole. Therefore, additional simula-
tions were needed to ensure that we could achieve the desired mass filtration with
our modified quadrupole. Though the fields produced by the modified QMF closely
resemble the fields of the ideal QMF, there must be a significant change in the way the
QMF is analyzed computationally. With the ideal quadrupole, the Mathieu equation
alone accurately describe ion motion. When we depart from the ideal case, we can no
longer rely on the equations, but must enter the realm of numerical calculation. This
shift required a significant increase in computational work which will be explored inthe next section.
2.3 Computer Modeling of the Quadrupole
In order to ensure that our modified quadrupole design would work as expected, I
simulated ion motion in quadrupole fields. This included solving differential equations
for the ideal quadrupole case and solving ion motion in fields generated from potential
grids.
I began my computer simulations by solving the Mathieu equation to produce
trajectories for ions in an ideal quadrupole. I used Matlabs ordinary differential
equation solver, ode45, to solve the Mathieu equation in each direction in time. Ex-
amples of ion trajectories are shown in Fig. 2.2.
The next step in simulation was the transition from solving ion motion in the idealquadrupole to solving ion motion in fields generated by a cylindrical quadrupole. The
first step in this process was to create a two-dimensional grid on which I could solve
Poissons equation
2V =
0(2.3)
-
7/29/2019 2008 Giraud Kellen Senior Thesis
26/79
CHAPTER 2. METHODS 18
0.05 0 0.050.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
(a) Fields in the x direction
0.05 0 0.05
0.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
(b) Fields in the y direction
Figure 2.6: Contour lines for the electric field values in the x and y directions in the
modified quadrupole. Dark lines represent low electric field values, while light lines
represent high values.
to calculate voltages at each point on the grid. I used the method of successive over-
relaxation (SOR)[14] (see Appendix B for Matlab code). This code produced fields
shown in Fig. 2.5.
After generating the voltage grids for the modified quadrupole, I solved for the
electric fields by taking the spatial gradient of the voltage grid. Unlike voltage, a scalar
quantity, electric field is a vector quantity and thus I needed to solve for the electric
field in both the x and y directions. Figure 2.6 shows contour lines for the values of
the electric field in the x and y directions. The electric fields gain extreme amplitudes
near the surfaces of the quadrupole rods, but near the center of the quadrupole the
change in electric field strength is nearly linear. Because of the symmetry of the
quadrupole, the electric fields in the y direction exactly parallel the behavior of the
fields in the x direction. The contours are not the electric field lines, and ions in these
fields travel in a direction perpendicular to the superposition of the two contours.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
27/79
CHAPTER 2. METHODS 19
Next I applied the electric field grids to a code to solve for ion motion in the
new fields. I put the electric fields into differential equations for the forces on the
ions, and solved the differential equations with ode45. After obtaining solutions
for a two-dimensional quadrupole, I generalized to a three-dimensional quadrupole
using three-dimensional grids for voltage and electric fields. With solutions for ion
trajectories in the three-dimensional case, I was able to compare the ideal quadrupole
to our modified quadrupole and make predictions about how we would need to build
and operate the QMF in our experiment.
Before I could calculate the motion of ions in the QMF using the electric fieldsin differential equations, I needed to have initial conditions in the equations. In my
simulations I used initial conditions based on data collected from experimentation of
the ion source. The ion source[8] ejects ions from a point on the surface of a sample.
The ions spray outward in a cone with most of their momentum being perpendicular
to the sample surface. Thus, the initial conditions of a particular ion can be expressed
with the parameters of initial position in the x and y directions, total initial velocity,
ratio of radial velocity to longitudinal velocity, and angle of initial motion with respect
to the x axis. In simulations, I converted these parameters to initial position and
velocity in the x, y, and z directions. See appendix B for Matlab code.
2.4 The Magnetic Field
In order to accurately represent the physical experiment, my simulations needed toincorporate the magnetic field generated by the ion trap (see Fig. 1.1). A conceptual
picture of the magnetic field lines emerging from the ion trap is presented in Fig.
2.7. The strength of the magnetic field is represented by the density of magnetic field
lines. The quadrupole and ion source share a concentric axis with the ion trap. On
-
7/29/2019 2008 Giraud Kellen Senior Thesis
28/79
CHAPTER 2. METHODS 20
IonTrap
MagneticField Lines
BzBr
Quadrupole
Figure 2.7: A conceptual schematic of magnetic field lines as they exit the ion trap-
ping region. The magnetic field line can be separated into radial and longitudinal
components, as shown here.
the axis of the trap the magnetic field points mostly in the longitudinal direction, and
grows weaker with distance away from the trap.
The values for the longitudinal component of the magnetic field, Bz in Fig. 2.7,
are plotted as a function of distance in Fig. 2.8(a). These values were obtained
by direct measurement. The magnetic field in the radial direction arises because
of the change in the field in the longitudinal direction. That is, any change in the
longitudinal magnetic field means the field escaped in the radial direction, so the field
in the radial direction is proportional to the derivative of the field in the longitudinal
direction. The radial field is also proportional to distance from the axis in the region
near the axis. Thus, the values of the magnetic field in the radial direction canbe obtained by taking the derivative of the field in the longitudinal direction and
multiplying by the distance from the axis. These values are shown in Fig. 2.8(b) for
a point 2.5 cm off axis.
In order to apply the magnetic field to the simulations, I created grids of magnetic
-
7/29/2019 2008 Giraud Kellen Senior Thesis
29/79
CHAPTER 2. METHODS 21
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Distance (m)
MagneticField(
T)
Start ofTrap
Location ofIon Source
QuadrupoleRegion
(a) Field in longitudinal direction
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80
0.005
0.01
0.015
0.02
0.025
0.03
0.035
Distance (m)
MagneticField(
T)
QuadrupoleRegion
Location ofIon Source
Start ofTrap
(b) Field in radial direction 2.5cm from axis
Figure 2.8: The magnetic field values in the longitudinal and radial directions as a
function of distance from a point slightly beyond the ion source into the ion trap.
The bold lines are the magnetic field values, and the labels show the locations of the
sample, the quadrupole, and the near side of the ion trap, as placed in the grid I used
in my simulations.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
30/79
CHAPTER 2. METHODS 22
field values in three directions with the same size as the grids of electric field values.
To create the grids for the magnetic field in the longitudinal direction, I assumed that
radial distance from the axis has a negligible effect on the value of the magnetic field.
That is, the value of the field in the longitudinal direction is the same on the axis as
it is at any point off of the axis.
After obtaining grids for the magnetic field, I simply incorporated the effect of the
magnetic fields in my previous simulations in accordance with the Lorentz equation
for electric and magnetic forces,
F = q( E+ v B) (2.4)
Once the magnetic field was incorporated into the simulation in this way, its effects
could be observed on the motion of the simulated ions.
In the simulations, the magnetic field has a negligible effect on charged particles
until they have penetrated about 20 cm into the QMF. At a some point in the
trajectory of an ion the influence of the magnetic field overpowers the influence of the
electric field, resulting in cyclotron motion. The transition from motion dominated
by the electric field to motion dominated by the magnetic field can be observed in Fig.
2.9. In motion in the magnetic field, the ion circles with a Larmor radius governed
by the equation
r =mvqB
(2.5)
while the frequency of this motion is
f =eB
2m(2.6)
Many QMFs serve the purpose of breaking the magnetic mirror effect.[10] The
magnetic mirror effect is that charged particles entering a region of high magnetic field
tend to reflect back in the opposite direction. This is because particles entering the
-
7/29/2019 2008 Giraud Kellen Senior Thesis
31/79
CHAPTER 2. METHODS 23
region with any component of velocity perpendicular to the direction of the magnetic
field lines begin cyclotron motion. When the particle enters cyclotron motion, the
kinetic energy of the particle converts from longitudinal energy that drives the particle
into the magnetic field to radial energy. If the initial perpendicular velocity of the
particle is great enough, its longitudinal momentum will become entirely converted
to radial momentum at some point, and the particle reverses direction to travel away
from the magnetic field. A quadrupole can break this effect by randomizing the
perpendicular velocity of the particle as it approaches the region of high magnetic
field.While the magnetic mirror effect is important in many experiments involving
charged particles entering regions of high magnetic field, we have determined that this
effect is of minimal importance in our experiment. This is because even without our
QMF, ions entering the magnetic field will not encounter the magnetic mirror effect
due to the geometry of our system. Ions that have enough perpendicular velocity
to become subject to the magnetic mirror effect hit the walls of our system before
having a chance to do so. Thus, in our experiment, we determined that it was safe
to ignore the magnetic mirror effect in building the QMF.
We found that the greatest influence that the magnetic field has on the function
of the QMF is the condition of resonance. When the cyclotron frequency (Eq. 2.6)
of an ion in the QMF is near the frequency of the AC voltage applied to the QMF,
resonance occurs and the ion is driven into circular motion with increasing radius.
When resonance occurs, there is a high probability of the ion being ejected from the
quadrupole.
For singly ionized 7Be in a magnetic field of 0.43 T (maximum magnetic field in
the trap), the cyclotron frequency, calculated using Eq. 2.6, is 941 kHz. However,
as the ion approaches the trap, it has a cyclotron frequency that is proportional
-
7/29/2019 2008 Giraud Kellen Senior Thesis
32/79
CHAPTER 2. METHODS 24
0.03 0.02 0.01 0 0.01 0.02 0.030.025
0.02
0.015
0.01
0.005
0
0.005
0.01
0.015
0.02
0.025
x (m)
y(m)
(a) 40 cm quadrupole.
0.03 0.02 0.01 0 0.01 0.02 0.030.025
0.02
0.015
0.01
0.005
0
0.005
0.01
0.015
0.02
0.025
x (m)
y(m)
(b) 44 cm quadrupole.
Figure 2.9: The figure at left shows ion motion in the x-y plane for a QMF 40 cm
long. This is the length that is shown in Fig. 2.8. Note that the QMF does not reach
the half-field value. The figure at right shows ion motion for a QMF 44 cm long.
At the far end of this QMF, the value of the magnetic field is 0.17 T, and resonance
occurs driving the ion out if the system.
to the magnetic field strength. For example, when the ion reaches the point where
the magnetic field is half its maximum value (at about 0.56 m in Fig. 2.8(a)), the
cyclotron frequency is 470 kHz. Since we desire to operate the quadrupole with an
AC frequency of about 550 kHz, the quadrupole must not penetrate to where the
magnetic field is at half-strength to avoid the effects of resonance.
This effect is illustrated in Fig. 2.9. When the QMF is 40 cm long (Fig. 2.9(a)),
it extends to where the magnetic field is 0.10 T. At this point the cyclotron radius of
7Be is 218 kHz, well below 550 kHz. Thus, resonance is avoided. With the same initial
conditions applied when the QMF is 44 cm long (Fig. 2.9(b)), resonance occurs. Here,
the QMF extends to where the magnetic field is 0.17 T and the cyclotron frequency
-
7/29/2019 2008 Giraud Kellen Senior Thesis
33/79
CHAPTER 2. METHODS 25
is 371 kHz, which is only a factor of 1.5 away from the frequency of the AC voltage.
As a result of these simulations, our quadrupole was built 40 cm long to avoid the
effects of resonance.
2.5 The Modified Stability Curve
One of the most important results of my simulations was knowledge about how mod-
ifying the geometry of the quadrupole changes how it functions. I used computer
simulations to test the stability of ions in the modified quadrupole. Based on these
tests, I empirically found where the stability curves lie for the modified QMF. The
result is shown in Fig. 2.10. As in seen in the figure, the modification to the geometry
of the quadrupole results in an extended region of stability to the right of the stable
region for the ideal quadrupole. This means that more AC and DC voltage is required
to operate the QMF in the region of greatest mass resolution. However, the peak of
the modified stability region is plagued by instability in the magnetic field. This is
because even stable ions gain large amplitudes away from the axis of the quadrupole
when it is operated at higher voltages. The region of instability in the magnetic field
is somewhat arbitrary, since it depends on the initial conditions of ions entering the
QMF.
To avoid the problem of instability in the magnetic field, it is recommended that
the QMF be operated at close to q = 0.65 and a = 0.175 in our experiment. This
works because the stable region for
7
Be is to the left of the stable regions for the otherisotopes in our experiment, as shown in Fig. 2.3.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
34/79
CHAPTER 2. METHODS 26
Figure 2.10: The final stability curve as empirically determined in my simulations.
Using a modified quadrupole extends the normal region of stability to the right.
However, the peak of the stable region is actually unstable in the magnetic field.
2.6 Quadrupole Fabrication
Based on the data from the simulations, a quadrupole was manufactured to fit the
criteria required for filtering out all ions except 7Be. Graduate student David Olson
and I worked together to design a system in the program SolidWorks that could
provide support for the ion source as well as incorporate the QMF into the system.
The necessary features of our design were that it include only materials that would
not compromise our ultra-high vacuum system, that the ion source and quadrupole
be electrically isolated from all other parts, and that the system maintain a stable
position when undergoing mechanical perturbations.
The final design is shown in Fig. 2.11. Two brackets clamp on the insulating ring
of the ion source. The clamp on the left in Fig. 2.11 has a hole in the top for a bolt to
extend out from to stabilize the structure against the inner ceiling of the system. The
support plates for the QMF are fabricated with as little material as possible to allow
for maximum transmission of gas molecules to the vacuum pump. The quadrupole
-
7/29/2019 2008 Giraud Kellen Senior Thesis
35/79
CHAPTER 2. METHODS 27
Figure 2.11: An image of the ion source and QMF with the support structure, as
designed in SolidWorks.
is electrically isolated from the support structure by glass beads. This is shown in
picture in Appendix D. Drawings of each of the pieces in Fig. 2.11 are included in
Appendix C.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
36/79
Chapter 3
Results and Conclusions
3.1 Comparison of Simulation and Experiment
In this chapter I will discuss the data that has been generated by the quadrupole since
it has been incorporated into the system that will measure the half-life of 7Be (see
Fig. 1.1). While most of the results in my project have been computer-generated, the
manufactured quadrupole has also generated a significant amount of data. While we
are far from understanding all of the data, we have been able to draw some conclusions
about the experiment.
The experiment has far greater complexity than my computer simulations. In my
simulations, I only modeled the motion of one ion at a time. I could always put the
desired voltages on the rods exactly how I wanted them. I ignored any influences of
any other pieces of equipment in the system. In my simulations, I had one ion with
a set of well-defined initial conditions, traversing a region of well-defined fields.
The actual experiment is far more complex than my simulations, which makes
the data difficult to analyze and compare with simulation. Part of the complexity
comes from the fact that the experimental data is produced by all of the pieces of
28
-
7/29/2019 2008 Giraud Kellen Senior Thesis
37/79
CHAPTER 3. RESULTS AND CONCLUSIONS 29
the experimental setup acting together, each with their own complex behavior. The
ion source is one of the complex pieces of equipment. When the ion source is fired it
produces a complicated signal of noise and charges that varies somewhat with each
firing. The quadrupole itself is complex in that it is sometimes difficult to specify
the voltages on the rods exactly. In our experiments to this point, we have noticed
inconsistencies in our application of both the AC and DC components of the voltage
that we apply to the QMF.
Another source of complexity is the charge collectors. A good deal of electric
circuitry is involved in obtaining the data signal from the collectors. The only thingthe charge collector can measure is charge. It is possible that some of the measured
charge is actually induced by stray electric fields produced by the ion source, and has
nothing to do with the amount of ions striking the charge collectors. In addition,
it is difficult to determine exactly when and how many ions hit the charge collector
because of the effects of secondary electrons. Secondary electrons are electrons that
get ejected from the charge collectors when high-energy ions strike them. As a result,
one ion is measured as being two or more charges. The ion source, the QMF itself,
and the charge collectors are each individually complex parts, and each must be
understood separately in order for the data to be understood as a whole.
The experiment is more complex than my simulations aside from the complexities
and inconsistencies of the equipment. Additional complication comes from the fact
that in the experiment billions of ions all interact together as a plasma. When the
density of plasma in a QMF is low enough, the analysis I used in my simulations
accurately predicts the behavior of ions in the quadrupole because the interactions
between ions is minimal. However, when the plasma has a high density, plasma
effects become important. Estimating that the quadrupole region has a volume of
about one liter and there are about 1010 ions in the QMF during the experiment,
-
7/29/2019 2008 Giraud Kellen Senior Thesis
38/79
CHAPTER 3. RESULTS AND CONCLUSIONS 30
we calculated that the density of the plasma in the quadrupole is about 10 7 ions per
cubic centimeter. This means that the Debye length of the plasma is about 0.25 cm.
The Debye length is the approximate depth to which electric fields can penetrate into
the plasma. This means that aside from equipment inconsistency, the QMF probably
still will not work as a mass filter as is because it cannot reach all of the ions that it
is supposed to filter.
While the experiment has much greater complexity than was included in my sim-
ulations, there are some things that can be done in the future to make the quadrupole
operate effectively in our experiment. One option is to somehow incorporate the com-plexities of the experiment into a simulation of the QMF. With more powerful and
accurate simulations, we can better understand how the quadrupole can be success-
fully operated. A particle-in-cell code might be used to simulate the behavior of the
ion plasma in the QMF. Monte Carlo simulations may give insight into how the ions
behave in the QMF.
Another option to adjust the quadrupole so that it works as a mass filter is to
modify the experiment so that it matches my simulations. It may be possible to
reduce the density of plasma by passing the ions through a screen near the ion source
that only allows a percentage of the ions to pass into the quadrupole. Once this was
accomplished, each individual part of the experiment would need to be understood
well enough that we could produce conditions exactly like the conditions of the sim-
ulation. This method has the downside of limiting the number of particles that are
allowed into the trap at one time. Since we want to trap as many7
Be ions as possible,
this method is not preferable.
The last option to obtain a working QMF is to modify it to apply to plasma
conditions. I will explain some potential procedures to do this later in this chapter.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
39/79
CHAPTER 3. RESULTS AND CONCLUSIONS 31
3.2 Materials in the Quadrupole
Stepping aside from the experimental complexities of our setup, I will show some of
the data and conclusions that have been generated in tests of the QMF. We have
discovered that the composition of the quadrupole pieces was part of the problem
with the performance of the original QMF. The entire structure of the quadrupole
and support system was originally constructed of aluminum. The aluminum rods
have since been replaced with copper rods.
Aluminum is perfectly suitable as an electrical conductor, and can hold voltages
that are needed in a QMF. However, aluminum has the problem that a layer of
aluminum oxide (Al2O3) approximately 35 nanometers thick forms on the its surface
if left in atmosphere for any significant amount of time. Aluminum oxide is an
electrical insulator, so any charge that builds up on the Al2O3 surface stays for an
appreciable amount of time, especially in ultra high vacuum conditions.
The problem with using aluminum for the quadrupole rods was discovered as we
tested our experiment repeatedly in succession. We found that after an initial test, wecould run the test again with the exact same parameters and entirely different results
were produced. On the initial test we often recorded on the order of 108 Coulombs
on the charge collector, but on the next test we would get almost no charge collected.
If we ran the test again after waiting a sufficient amount of time (5-10 minutes), an
amount of charge comparable to the original amount would again be collected. We
attributed this effect to a surface charge building up somewhere in the quadrupole
system. We believe this occurred as ions struck parts of the quadrupole and built up
a surface charge on the Al2O3. The most likely spot for a charge buildup to occur is
at the support plate at the far end of the quadrupole (see Appendix C for drawings).
I modeled the effect of the surface charge on the support plate, and found that the
-
7/29/2019 2008 Giraud Kellen Senior Thesis
40/79
CHAPTER 3. RESULTS AND CONCLUSIONS 32
x
y
0.05 0 0.050.04
0.03
0.02
0.01
0
0.01
0.02
0.03
0.04
Figure 3.1: A contour plot of voltages produced by a support plate (shown at far
right in Fig. 2.11) charged with 1 109 Coulombs. Light lines are higher voltages
(with a maximum of 250 V) while dark lines are lower voltages (with a minimum of
150 V).
simulation confirms our hypothesis about charge buildup. Figure 3.1 shows a contour
plot of voltage caused by a charge buildup of one nanocoulomb on a simulated support
plate. Such a charged plate would cause fields that are capable of ejecting all ions
from the QMF. This demonstrates the debilitating effect of using an insulator capable
of holding a surface charge in the quadrupole.
To fix this problem, we replaced the aluminum rods and support plate on the
far end of the quadrupole with parts of the exact same dimensions fabricated out of
copper. Copper can form a copper oxide layer, but this layer is much thinner than
an aluminum oxide layer and is a semi-conductor rather than an insulator. Surface
charge that may build up on the copper oxide decays within a few milliseconds. By
using copper parts in the QMF, we can run repeated tests without charge buildup
causing distorted fields.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
41/79
CHAPTER 3. RESULTS AND CONCLUSIONS 33
3.3 The Effect of the Magnetic Field
One obvious result of our tests was that the magnetic field has a greater effect on the
ions moving through the quadrupole than was expected from my simulations. Figure
3.2 shows the charge collected at the different components of the charge collector
for tests with the magnetic field off (Fig. 3.2(a)) and on (Fig. 3.2(b)). When the
magnetic field is off, a total of only about 6 nanocoulombs of charge is collected among
the four charge collectors, concentrated on the outer edge of the collectors. When
the magnetic field is on, a total of about 150 nanocoulombs of charge is collected,
concentrated in the center of the collectors. Note the use of the prefix nano. It is
essential in any physics research to include this prefix where possible. Physics projects
that include the prefix nano are much more likely to get funding and recognition.
The data in Fig. 3.2 were taken with the magnetic field value at 40% strength.
When the magnetic field is on at full strength, approximately 100 times more charge
is collected than a test without the magnetic field. Whether the magnetic field is
on or off has far more influence on whether ions pass through the quadrupole thanwhether the quadrupole is turned on or not.
The results generated from tests on the actual experimental setup are very different
from what we expected to see based on data generated in my simulations of the
quadrupole. Based on simulations of the quadrupole, the amount of ions that get
through the quadrupole should not be influenced much by whether the magnetic field
is on or off. We may be able to fix this problem by placing the start of the QMF
farther back in the system, away from the strong magnetic field. We may need to
make another measurement of the magnetic field to ensure that the values are what
I have applied in my simulations.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
42/79
CHAPTER 3. RESULTS AND CONCLUSIONS 34
0 20 40 60 80 100 120
0
0.5
1
1.5
2
2.5
3
3.5
4
Time (s)
AccumulatedCharge(nC)
Trap RingInner Collector
Middle CollectorOuter Collector
(a) No magnetic field.
0 20 40 60 80 100 12020
0
20
40
60
80
100
120
140
160
Time (s)
AccumulatedCharge(nC)
Trap RingInner CollectorMiddle CollectorOuter Collector
(b) With magnetic field reaching a maxi-
mum value of 0.17 T.
Figure 3.2: The charge collected at the charge collector in time changes drastically
based on whether or not the magnetic field is on.
3.4 Reduction of Fringing Fields
Another possible source of error in the experimental quadrupole is the effect of fringing
fields produced by the ion source. During ion source operation, high voltage arcing
releases ions from the sample. Electric fields produced in this process can leak into the
QMF causing distortion of the desired fields and reduced quadrupole performance.
The effects of this occurrence are extremely difficult to predict, the ion source yields
a slightly different pattern of high voltage release each time it is fired.
Figure 3.3 shows the effect of a DC voltage being applied to the ion source while
in operation. When the front piece of the ion source is at a voltage of only 15 V,
the path of the ion is significantly affected, and even more so when 30 V is applied.
While this is only a simulation of the influence of voltage on the ion source, it shows
that the voltage can make a significant difference in the function of the QMF.
It may be possible to reduce the effect of fringing fields by installing a grounded
-
7/29/2019 2008 Giraud Kellen Senior Thesis
43/79
CHAPTER 3. RESULTS AND CONCLUSIONS 35
0.03 0.02 0.01 0 0.01 0.02 0.030.03
0.02
0.01
0
0.01
0.02
0.03
(a) No voltage on the ion source.
0.03 0.02 0.01 0 0.01 0.02 0.030.03
0.02
0.01
0
0.01
0.02
0.03
(b) 15 volts on the ion source.
0.03 0.02 0.01 0 0.01 0.02 0.030.03
0.02
0.01
0
0.01
0.02
0.03
(c) 30 volts on the ion source.
Figure 3.3: Plots of the first 7 microseconds of ions exiting the ion source with initial
conditions x0 = 1 mm, y0 = 1 mm, vr/vz = 0.05. Changing the voltage applied to
the ion source significantly changes the trajectories.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
44/79
CHAPTER 3. RESULTS AND CONCLUSIONS 36
copper mesh across the hole in the ground plate through which ions pass from the
ion source into the QMF. An additional effect of this approach would be a slight
reduction in ion current, as some would strike the mesh.
3.5 Potential QMF Modifications
As mentioned in section 3.1, though the QMF has produced much data and some
conclusions have been drawn, we are still far from being able to use the quadrupole in
our experiment to isolate a 7Be ion plasma. It is my purpose in this section to outline
possible ways of improving the function of the quadrupole so that mass resolution
can be achieved. I will cite the research that other groups have performed employing
the use of a QMF. I will not delve into the full details of the research, but leave it to
the reader to consult the sources mentioned.
Wexler [15](experimental) and Wagner [16](theory) used a quadrupole in an ex-
periment with a cesium-uranium-fluoride plasma. Their work may be useful in our
experiment because their plasma also had a high density, so they modeled the function
of a QMF with a high density plasma. To improve the function of the quadrupole,
Wexler and Wagner used a superimposed dipole field in conjunction with the nor-
mal quadrupole fields. It may be possible for something similar to be done in our
experiment.
Konenkov et al. added auxiliary quadrupole excitation to improve performance
of their QMF. [17] They used an additional AC voltage applied to the quadrupoleto alter the shape of the stability curves, obtaining very precise regions of stability.
Takahashi et al. used a Bessel-box type energy analyzer in a quadrupole. By altering
the way that the ions enter the QMF from the ion source, they eliminated distracting
noise from their experiment. [18] Tsukakoshi et al. used a zero-method control circuit
-
7/29/2019 2008 Giraud Kellen Senior Thesis
45/79
CHAPTER 3. RESULTS AND CONCLUSIONS 37
to apply voltages to their QMF. [19] Using this circuit, the voltages applied to the
quadrupole could be specified with increased accuracy.
The important result of all this research is that it is possible to use a QMF in
experiments similar to ours to obtain the desired results. Using previous research as
a guide, it may be possible for us to modify our quadrupole to enable it to isolate a
7Be plasma.
3.6 Conclusion
While the QMF has been simulated and manufactured, we are still far from under-
standing it. We cannot yet operate the quadrupole accurately enough to use it in the
experiment of measuring the half-life of ionized 7Be.
In future research of the QMF, it will be important to gain a good understanding
of each component that contributes to the function of the quadrupole. The ion source
must be analyzed so that its inconsistencies are understood. The DC and AC com-
ponents of the voltages on the quadrupole must be made so that they can be applied
without error. The magnetic field must be measured accurately and completely, and
its effects on ions independent of the QMF must be understood.
There are a number of ways that the QMF may be understood enough to be used in
the experiment. It may be possible to create simulations that more accurately reflect
the behavior of the experimental quadrupole. It may also be possible to modify the
experiment so that it more closely matches the conditions of my simulations. Onemethod to accomplish this is reducing the current from the ion source to reduce the
density of plasma in the quadrupole. Another option is to alter the function of the
QMF as outlined in the previous section to make mass filtration possible.
It is hoped that understanding of the quadrupole will be obtained and the quadrupole
-
7/29/2019 2008 Giraud Kellen Senior Thesis
46/79
CHAPTER 3. RESULTS AND CONCLUSIONS 38
will be an essential part of the system that will measure the half-life of ionized 7Be.
The measurement of this half-life will yield a better understanding of the variable
decay rate of this isotope and the studies of which it is a part.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
47/79
Bibliography
[1] C. W. Johnson, The Fate of7Be in the Sun, Astrophys. J. 392, 320327 (1992).
[2] M. Yoshimori, Production and behavior of beryllium-7 radionuclide in the upper
atmosphere, Adv. Space Res. 36, 922926 (2005).
[3] D. E. Walling, Q. He, and W. Blake, Use of 7Be and 137Cs measurements
to document short- and medium-term rates of water-induced soil erosion on
agricultural land, Water Resour. Res. 35, 38653874 (1999).
[4] D. Prime, V. Smith, and F. M. Ashworth, Beryllium-7 in sheep, J. Radiol.
Prot. 8, 239 (1988).
[5] S. Wexler, E. K. Parks, C. E. Young, and R. A. Bennett, Enhanced Electron-
Capture Decay Rate of 7Be Encapsulated in C60 Cages, Phys. Rev. Lett. 93,
112501 (2004).
[6] C. Huh, Dependence of the decay rate of7Be on chemical forms, Earth Planet.
Sci. Lett. 171, 352 (2004).
[7] L. Liu and C. Huh, Effect of pressure on the decay rate of 7Be, Earth Planet.
Sci. Lett. 180, 163 (2000).
[8] D. Olson, Development of a MEVVA Based Beryllium-7 Plasma Source, M.S.
thesis (Brigham Young University, Provo, Utah, 2007).
39
-
7/29/2019 2008 Giraud Kellen Senior Thesis
48/79
BIBLIOGRAPHY 40
[9] D. Olson, Construction and Characterization of a Specialized Malmberg-
Penning Trap for Confinement of a Nonneutral Radionuclide Plasma, Under-
graduate senior thesis (Brigham Young University, Provo, Utah, 2005).
[10] J. Rober T. McIver, Trajectory Calculations for Axial Injection of Ions into
a Magnetic Field: Overcoming the Magnetic Mirror Effect with an R.F.
Quadrupole Lens, Int. J. Mass Spectrom. Ion Processes 98, 3550 (1990).
[11] R. March and R. Hughes, Quadrupole Storage Mass Spectrometry (Wiley, New
York, 1988).
[12] P. H. Dawson and N. R. Whetten, in Advances in Electronics and Electron
Physics, L. Marton, ed., (Academic, New York, 1969).
[13] D. R. Denison, Operating parameters of a quadrupole in a grounded cylindrical
housing, J. Vac. Sci. Technol. 8, 266269 (1971).
[14] R. L. Spencer, Computational Physics 430: Partial Differential Equations
(Brigham Young University, Department of Physics and Astronomy, 2006).
[15] S. Wexler, E. K. Parks, C. E. Young, and R. A. Bennett, Behavior of Cs-UF ion-
pair plasmas in radiofrequency quadrupole-dipole fields Experiment, J. Appl.
Phys. 54, 17301754 (1983).
[16] A. Wagner, Behavior of Cs-UF ion-pair plasmas in radiofrequency quadrupole-
dipole fields Theory, J. Appl. Phys. 54, 17551780 (1983).
[17] N. Konenkov, L. Cousins, V. Baranov, and M. Sudakov, Quadrupole mass filter
operation with auxiliary quadrupolar excitation: theory and experiment, Int.
J. Mass Spectrom. Ion Processes 208, 1727 (2001).
-
7/29/2019 2008 Giraud Kellen Senior Thesis
49/79
BIBLIOGRAPHY 41
[18] N. Takahashi, T. Hayashi, H. Akimichi, and Y. Tuzi, Development of the
quadrupole mass spectrometer with the BesselBox type energy analyzer: Func-
tion of the energy analyzer in the partial pressure measurements, J. Vac. Sci.
Technol. A 19, 16881692 (2001).
[19] O. Tsukakoshi, T. Hayashi, K. Yamamuro, and M. Nakamura, Development of
a high precision quadrupole mass filter using the zero-method control circuit,
Rev. Sci. Instrum. 71, 13321336 (2000).
-
7/29/2019 2008 Giraud Kellen Senior Thesis
50/79
Appendix A
The Mathieu Equation
The Mathieu equation is of great importance to the function of the quadrupole mass
filter. Knowledge about how to operate the QMF is obtained from this equation.
Much of this appendix is a summary of the analysis of March and Hughes. [11]
To arrive at the Mathieu equation for ion motion in a quadrupole field, we start
with the potential experienced by an ion in a quadrupole field. The potential is
= 0
r20(x2 + y2 + z2) (A.1)
in which , , and are weighting constants for the respective coordinates, 0 is the
applied electric potential, and r0 is the inner radius of the quadrupole as shown in
Fig. 2.4. The applied potential on the QMF is
0 = U V cos(t) (A.2)
When the potential experienced by the ion (Eq. A.1) is put in Poissons equation(Eq. 2.3) with = 0, the condition
+ + = 0 (A.3)
is obtained. In the two-dimensional quadrupole case, the potential in the z direction
is zero. Thus, = 0 and the limitations on the weighting constants can be satisfied
42
-
7/29/2019 2008 Giraud Kellen Senior Thesis
51/79
APPENDIX A. THE MATHIEU EQUATION 43
if
= = 1 (A.4)
Thus, the field expression becomes
(x, y) =0r20
(x2 y2) (A.5)
We can put the expression for 0 (Eq. A.2) into Eq. A.5 and differentiate with respect
to x and y to obtain a potential gradient in each direction,
x
=2x
r2
0
(U V cost) (A.6)
y=
2y
r20(U V cost) (A.7)
The electrical force on the ion in each direction is the potential gradient multiplied
by the charge. Force also equals mass times acceleration. Equating the expressions
for the forces, we obtain
Fx = md2x
dt2= 2ex
r20(U V cost) (A.8)
Fy = md2y
dt2=
2ey
r20(U V cost) (A.9)
Thus, the equations of motion for a singly charged positive ion in a two-dimensional
quadrupole field are
d2x
dt2 +
2e
mr20 (U V cost) x = 0 (A.10)
d2y
dt2+
2e
mr20(U V cost) y = 0 (A.11)
d2z
dt2= 0 (A.12)
-
7/29/2019 2008 Giraud Kellen Senior Thesis
52/79
APPENDIX A. THE MATHIEU EQUATION 44
The equations in the x and y directions resemble the second-order linear differ-
ential equation known as the Mathieu equation. Mathieu originally developed this
equation while investigating the motions of a vibrating membrane. The canonical
form of the equation is
d2u
d2+ (au 2qucos2) u = 0 (A.13)
with u representing one of the coordinate axes. If we change variables in Eq. A.10
and Eq. A.11 so that
=t
2(A.14)
we can obtain
d2x
d2=
8exU
mr202
+8exV cos2
mr202
(A.15)
and the same for the y direction. If we define ax and qx such that
ax =8eU
mr202
(A.16)
and
qx =4eV
mr202
(A.17)
we obtain
d2x
d2+ (ax 2qxcos2) x = 0 (A.18)
and the same equation in the y direction. Note that the values for a and q are the
same as those in Eq. 2.1 and Eq. 2.2. Equation A.18 is the canonical form of the
Mathieu equation. Thus, we can consult the knowledge of this historically importantequation which has been studied extensively to learn about the function of the QMF.
The stable, bounded solutions for the Mathieu equation are well known. Stable
solutions in the x direction fit under the curve
ax0 = 1 q1
8q2 +
1
64q3 +
1
1536q4 + ... (A.19)
-
7/29/2019 2008 Giraud Kellen Senior Thesis
53/79
APPENDIX A. THE MATHIEU EQUATION 45
This curve, plotted for a = 0.7 to a = 0.9, is on the right in Fig. 2.1. Meanwhile,
stable solutions in the y direction fit under the curve
ay0 =1
2q2
7
128q4 +
29
2304q6 + ... (A.20)
This curve, plotted for a = 0 to a = 0.7, is on the left in Fig. 2.1.
The stability curves for the ideal QMF come from the analysis of the well-known
Mathieu equation.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
54/79
Appendix B
Matlab Code
My project required a significant amount of coding in Matlab. Only the mostimportant programs from my simulations are included here. With all of these files ina directory, ion motion in our quadrupole field can be simulated.
B.1 Solving the Voltage Grid
This program creates a grid of voltages for a quadrupole with rods of a specifiedradius and length.
%Solve LaPlaces equation to obtain a grid of voltage values
clear;
close all;
%Input the grid points for the three-dimensional grid
%Include more points in the z direction because it is longer
N=25;
Nx=N;
Ny=N;
Nz=3*Nx;
%Define the radii of the necessary parameters
%In our QMF, the central region has radius 2cm, and the radius of the
%rods is half of that
r0=0.02;
r1=.5*r0;
l0=0.1; %The length from z=0 to the beginning of the quadrupole
l=.4; %The length of the quadrupole
lengthz=.567; %The length between the ion source and the ion trap
46
-
7/29/2019 2008 Giraud Kellen Senior Thesis
55/79
APPENDIX B. MATLAB CODE 47
ltot=lengthz+.3; %Add 30 cm for wiggle room for computation
%Define the length (diameter) of computation
Lx=.1; %Length in x of the computation regionLy=.1; %Length in y of the computation region
Lz=ltot; %Length in z of the computation region
% Define the grids
dx=Lx/(Nx-1); %X-grid spacing
dy=Ly/(Ny-1); %Y-grid spacing
dz=Lz/(Nz-1); %Z-grid spacing
x=(0:dx:Lx)-.5*Lx; %X-grid, centered at x=0
y=(0:dy:Ly)-.5*Ly; %Y-grid, centered at y=0
z=(0:dz:Lz); %Z-grid, starting at one end of the quadrupole
%Determine the approximate best omega: parameter in SOR
omega=1.72;
%Set the error criterion
errortest=1e-6;
%Define the voltages
U1=1; %Direct current potential
P1=U1; %Potential for two of the poles
P2=-U1; %Potential for the other two poles
%Define boundary conditions%Put the boundary conditions in a matrix U
%Grounded around the outside edge
U=zeros(Nx,Ny,Nz);
%Define the mask
mask=zeros(Nx,Ny,Nz);
%Make the grounded border around the outside
for i=1:Nx
for j=1:Ny
r=sqrt(x(i)^2+y(j)^2);
if r
-
7/29/2019 2008 Giraud Kellen Senior Thesis
56/79
APPENDIX B. MATLAB CODE 48
%Start with the left pole
%The center or the pole is r0 to the left
%of the center of the region
for i=1:Nxfor j=1:Ny
r=sqrt((x(i)+(r0+r1))^2+y(j)^2);
if r
-
7/29/2019 2008 Giraud Kellen Senior Thesis
57/79
APPENDIX B. MATLAB CODE 49
U(:,:,Nz)=0;
mask(:,:,1)=0;
mask(:,:,Nz)=0;
%The quadrupole itself is only defined at the center of the region of
%computation: we are solving the fields for free space outside the
%quadrupole in the z direction
%Thus, we can limit U to be zero anywhere outside the middle 30cm of the
%computation region
%First, define the limits in the index region
%Also fix the mask so that it iterates all the regions without poles
Nz1=l0;
Nz2=l0+l;
for n=1:Nz
if z(n)Nz2
U(:,:,n)=0;
mask(:,:,n)=1;
end
end
%Define the scale of the voltage, to be used to check the error in the
%iterationsUscale=max(max(max(abs(U))));
%Now plug the initial condition matrix U into V
%to be manipulated by iteration to determine the fields
V=U;
%Set a maximum iteration count
Niter=Nx*Ny*Nz;
%Set factors used repeatedly in the algorithm
facyz=dy^2*dz^2;facxz=dx^2*dz^2;
facxy=dx^2*dy^2;
for n=1:Niter
err(n)=0; %initialize the error at iteration n to zero
for i=2:(Nx-1) %Loop over interior points only
for j=2:(Ny-1)
-
7/29/2019 2008 Giraud Kellen Senior Thesis
58/79
APPENDIX B. MATLAB CODE 50
for k=2:(Nz-1)
if (mask(i,j,k))==1
rhs=(facyz*(V(i+1,j,k)+V(i-1,j,k))+facxz*...
(V(i,j+1,k)+V(i,j-1,k))+facxy*...(V(i,j,k+1)+V(i,j,k-1)))/...
(2*(facyz+facxz+facxy));
err(n)=max(err(n),abs(V(i,j,k)-rhs)/Uscale);
V(i,j,k)=omega*rhs+(1-omega)*V(i,j,k);
end
end
end
end
if (abs(err(n)) < errortest)
fprintf(After %g iterations, error= %g\n,n,err(n));
disp(Desired accuracy achieved; breaking out of loop);
break;
end
end
%make a surface plot
figure
surf(x,y,V(:,:,round(Nz/2)));
xlabel(x);ylabel(y);
title(Fields in the middle of the quadrupole)
%Make an animation of the fields as the z direction moves past the
%ends of the quadrupolefor m=1:Nz
if z(m)>(l0+l-l0/2)
break
end
end
for s=1:Nz
if z(s)>(l0+l+l0/2)
break
end
end
figurefor n=m:s
surf(x,y,V(:,:,n));
zlim([-1 1])
p=z(n)-(.3+3*(r0+2*r1));
title([Fields at ,num2str(p), outside the quadrupole])
pause
end
-
7/29/2019 2008 Giraud Kellen Senior Thesis
59/79
APPENDIX B. MATLAB CODE 51
%Display the maximum voltage at three points outside the quadrupole
v1=max(max(V(:,:,s-3)));
q=100*(z(s-3)-(z(s)+z(m))/2);fprintf(Voltage %1.4gcm outside the quadrupole - %1.4g \n,q,v1)
v1=max(max(V(:,:,s)));
q=100*(z(s)-(z(s)+z(m))/2);
fprintf(Voltage %1.4gcm outside the quadrupole - %1.4g \n,q,v1)
v1=max(max(V(:,:,s+3)));
q=100*(z(s+3)-(z(s)+z(m))/2);
fprintf(Voltage %1.4gcm outside the quadrupole - %1.4g \n,q,v1)
%Make a one-dimensional cross section of the voltage
vt1=round(Nx/4);
vt2=round(Ny/2);
Vline=reshape(V(vt1,vt2,:),1,Nz);
figure
plot(z,Vline)
%Make a slice across the quadrupole
Vslice=V((Nx/2)+.5,:,round(Nz/2));
figure
plot(x,Vslice,b-);
hold on
plot(r0,-1,r*)
plot(r0+dx,-1,r*)
plot(-r0,-1,r*)plot(-r0-dx,-1,r*)
%Define the error in the radius measurement
for n2=round(Nx/6):round(Nx/2)
if Vslice(n2)>-.99
break
end
end
xspecial=x(n2-1);
diff=abs(r0+xspecial);
%Save the data to be loaded in setup.m to solve for electric fields
save potential3coarse V x y z diff
B.2 Solving the Electric Fields
This program is to be run after the previous program to solve for the electric fieldsin the quadrupole.
-
7/29/2019 2008 Giraud Kellen Senior Thesis
60/79
APPENDIX B. MATLAB CODE 52
%Calculate the electric field from the potential in the x and y directions
%and load it into matrices Ex and Ey
clear;
close all;
load potential3coarse
xpot=x;
ypot=y;
zpot=z;
%Establish matrices for Ex, Ey, and Ez
Ex=zeros(length(x),length(y),length(z));
Ey=zeros(length(x),length(y),length(z));
Ez=zeros(length(x),length(y),length(z));
for n=2:length(x)-1
for m=2:length(y)-1
for k=2:length(z)-1
Ex(n,m,k)=(V(n-1,m,k)-V(n+1,m,k))/(2*(x(n+1)-x(n)));
Ey(n,m,k)=(V(n,m-1,k)-V(n,m+1,k))/(2*(y(m+1)-y(m)));
Ez(n,m,k)=(V(n,m,k-1)-V(n,m,k+1))/(2*(z(k+1)-z(k)));
end
end
end
%Save the electric fields
save setupvar Ex Ey Ez xpot ypot zpot diff
B.3 The Magnetic Field Values
This program gives the values for the magnetic field which were measured directlyfrom the experimental system.
% This will define the axial magnetic field and necessary info to
% calculate the axial and radial field (near the axis)
global Zvalues Bzvalues dBzvalues;
% These values are the Bz calculated with a fully-corrected magnet using
% the experimental values for the flattest central field.
Zvalues=0:.01:1.1;
Bzvalues=[0.4295466304,0.4295662094,0.4295963200,0.4296345496,...
0.4296763567,0.4297144393,0.4297378389,0.4297307893,...
-
7/29/2019 2008 Giraud Kellen Senior Thesis
61/79
APPENDIX B. MATLAB CODE 53
0.4296713377,0.4295298046,0.4292671812,0.4288336081,...
0.4281670866,0.4271925686,0.4258215344,0.4239521532,...
0.4214701565,0.4182506079,0.4141605817,0.4090617099,...
0.4028094744,0.3952456023,0.3861851524,0.3754072203,...0.3626580838,0.3476710715,0.3302084790,0.3101361090,...
0.2875384866,0.2628537666,0.2369436230,0.2109777410,...
0.1861265663,0.1632540784,0.1428023102,0.1248611022,...
0.1093023979,0.9589390572e-1,0.8437070998e-1,0.7447322880e-1,...
0.6596446198e-1,0.5863600761e-1,0.5230847743e-1,...
0.4682936807e-1,0.4206996264e-1,0.3792204876e-1,...
0.3429481525e-1,0.3111207999e-1,0.2830988893e-1,...
0.2583447768e-1,0.2364055984e-1,0.2168990148e-1,...
0.1995013867e-1,0.1839380060e-1,0.1699750108e-1,...
0.1574127219e-1,0.1460801272e-1,0.1358303204e-1,...
0.1265367221e-1,0.1180899351e-1,0.1103951398e-1,...
0.1033699044e-1,0.9694236812e-2,0.9104970776e-2,...
0.8563685486e-2,0.8065540777e-2,0.7606271621e-2,...
0.7182110507e-2,0.6789721112e-2,0.6426142183e-2,...
0.6088739303e-2,0.5775163515e-2,0.5483316050e-2,...
0.5211317625e-2,0.4957482239e-2,0.4720294228e-2,...
0.4498388514e-2,0.4290533437e-2,0.4095615636e-2,...
0.3912627056e-2,0.3740653115e-2,0.3578863421e-2,...
0.3426501954e-2,0.3282880101e-2,0.3147369297e-2,...
0.3019394995e-2,0.2898431358e-2,0.2783996687e-2,...
0.2675648590e-2,0.2572980875e-2,0.2475619731e-2,...
0.2383220755e-2,0.2295466419e-2,0.2212063618e-2,...
0.2132741446e-2,0.2057249063e-2,0.1985354278e-2,...0.1916841724e-2,0.1851511590e-2,0.1789178204e-2,...
0.1729668832e-2,0.1672822884e-2,0.1618490556e-2,...
0.1566532107e-2,0.1516817414e-2,0.1469224744e-2,...
0.1423640175e-2,0.1379957460e-2,0.1338076844e-2,...
0.1297904846e-2,0.1259354042e-2];
NBvalues=length(Zvalues);
dBzvalues(2:NBvalues-1)=(Bzvalues(3:NBvalues)-Bzvalues(1:NBvalues-2))./...
(Zvalues(3:NBvalues)-Zvalues(1:NBvalues-2));
dBzvalues(1)=3*dBzvalues(2)-3*dBzvalues(3)+dBzvalues(4);
dBzvalues(NBvalues)=3*dBzvalues(NBvalues-1)-3*dBzvalues(NBvalues-2)+...
dBzvalues(NBvalues-3);
plot(Zvalues,Bzvalues);
figure
-
7/29/2019 2008 Giraud Kellen Senior Thesis
62/79
APPENDIX B. MATLAB CODE 54
plot(Zvalues,dBzvalues);
save Bvalues Bzvalues dBzvalues Zvalues
B.4 The Magnetic Field Grid
This program puts the magnetic field values into a grid that is the same size as thegrid for the electric field.
clear;
close all;
%Load the values for the magnetic fields from Binit.m
load Bvalues
%Also load the variables from the electric field solver so we can get
%similar matrices to compare with the magnetic field matrices
load setupvar
%Start by changing the magnetic field values from Bvalues.m to values that
%correlate with the quadrupole
l0=0.1; %The location of the mouth of the quadrupole
ltot=0.867; %The length of the computation region
N=length(zpot);
dz=ltot/(N-1);
Bzpot=0:dz:ltot;
%Interpolate the magnetic fields onto the values that we need them for and
%plot the new values
Bzvaluesnew=interp1(Zvalues,Bzvalues,Bzpot);
subplot(2,2,1)
plot(Bzpot,Bzvaluesnew)
dBzvaluesnew=interp1(Zvalues,dBzvalues,Bzpot);
subplot(2,2,2)
plot(Bzpot,dBzvaluesnew)
%Reverse the direction of the magnetic fields and make the derivatives
%negativezB=Bzvaluesnew(N:-1:1);
dzB=-dBzvaluesnew(N:-1:1);
subplot(2,2,3)
plot(Bzpot,zB)
subplot(2,2,4)
plot(Bzpot,dzB)
%Build matrices for the magnetic field values
top related